Dynamics of the DEAD-box ATPase Prp5 RecA-like domains provide a conformational switch during spliceosome assembly

Abstract

The DEAD-box family of proteins are ATP-dependent, RNA-binding proteins implicated in many aspects of RNA metabolism. Pre-mRNA splicing in eukaryotes requires three DEAD-box ATPases (Prp5, Prp28 and Sub2), the molecular mechanisms of which are poorly understood. Here, we use single molecule FRET (smFRET) to study the conformational dynamics of yeast Prp5. Prp5 is essential for stable association of the U2 snRNP with the intron branch site (BS) sequence during spliceosome assembly. Our data show that the Prp5 RecA-like domains undergo a large conformational rearrangement only in response to binding of both ATP and RNA. Mutations in Prp5 impact the fidelity of BS recognition and change the conformational dynamics of the RecA-like domains. We propose that BS recognition during spliceosome assembly involves a set of coordinated conformational switches among U2 snRNP components. Spontaneous toggling of Prp5 into a stable, open conformation may be important for its release from U2 and to prevent competition between Prp5 re-binding and subsequent steps in spliceosome assembly.

Figure 1.

A Variant for Exploring the Conformational Dynamics of the Prp5 RecA-like Domains. (A) Structural model for conformational changes of Prp5 upon binding of ATP and RNA with predicted distances between sites of fluorophore incorporation in Prp5^core noted (open structure from 4LK2.pdb; closed model courtesy of Dr Yong-Zhen Xu, Wuhan U.). (B) Mutations incorporated into Prp5 proteins and domain architecture relative to the structures shown in panel (A). (C) Prp5 mutants support yeast viability in the absence of plasmid-borne wildtype PRP5 (+FOA) at 30°C. (D) Addition of poly(A) RNA stimulates the ATPase activity of fluorophore-labeled Prp5^core. ATPase assays were carried out in triplicate and the shown value represents the average ±SD.

Figure 2.

Conformational Switching of Prp5^core Requires ATP and RNA. (A) Schematic of the smFRET assay. (B) Top: Cy3 (green) and Cy5 (red) fluorescence trajectories from a single molecule of Prp5^core. Bottom: Calculated E_FRET from the shown trajectories. (C) Histogram of E_FRET values obtained for Prp5^core (grey boxes) and a fit of the data (blue line). (D) Data from a single molecule of Prp5^core as in panel (B) except in the presence of poly(A) RNA and ATP. (E) Histogram of E_FRET values obtained for Prp5^core in the presence of poly(A) RNA and ATP as in panel (C) and a fit of the data (red line). n and m represent the number of individual molecules and E_FRET data points, respectively, used to construct the histogram. Data fit results are listed in Supplementary Table S4.

Figure 3.

The Prp5-E235A substitution increases splicing fidelity in yeast and conformational dynamics of Prp5^core. (A) Schematic of the ACT1-CUP1 reporter pre-mRNA and assay. The BP-A is underlined, and positions of BS mutations are noted. (B–D) Results from Cu²⁺ growth assays of strains containing the indicated Prp5 mutations and either the consensus, U257C, or A258U ACT1-CUP1 reporter plasmids. Results are the average of quadruplicate experiments ± the standard deviation. (E, F) Representative E_FRET trajectories for two molecules of Prp5^core-E235A (G) Histogram of E_FRET values obtained for Prp5^core-E235A (grey boxes) and a fit of the data (red line). Blue line is from Figure 2C. (H) Representative calculated E_FRET trajectory for a single molecule of Prp5^core-E235A in the presence of poly(A) RNA and ATP. (I) Histogram of E_FRET values obtained in the presence of poly(A) RNA and ATP as in panel (G). n and m represent the number of individual molecules and E_FRET data points, respectively, used to construct the histogram. Data fit results are listed in Supplementary Table S4.

Figure 4.

Impact of E235A and SAT-motif Mutations on Prp5^core Dynamics. (A–C) Histograms of E_FRET values obtained for Prp5^core-GAG (A), -TAG (B) and -GAR (C) in the presence of poly(A) RNA and ATP (gray boxes). n and m represent the number of individual molecules and E_FRET data points, respectively, used to construct the histogram. (D) Calculated K_close values and ATPase rates in the presence of poly(A) RNA obtained for Prp5^core and the indicated mutants. Histogram fit results, calculated lifetimes, and K_close values are listed in Supplementary Table S5. ATPase assays were carried out in triplicate and the shown value represents the average ±SD.

Figure 5.

Potential Role for Prp5 Conformational Switching during Spliceosome Assembly. Release of Prp5 from U2 could be accompanied by spontaneous transition from the closed to the open conformation of the Prp5 RecA domains. Since the open conformation is stable in the absence of both RNA and ATP, this could render Prp5 release effectively irreversible (red forward arrow) and facilitate subsequent tri-snRNP binding. For simplicity, we have shown Prp5 and the tri-snRNP with overlapping binding sites on U2. While it is known that Prp5 blocks tri-snRNP addition (13), a structure of Prp5 bound to U2 has not yet been determined, nor is it known if the U2 snRNP undergoes conformational change after Prp5 release.